Ensuring the impeccable quality of integrated circuits (ICs) is of utmost importance to the customers. However, the increasing scaling of advanced nanoscale circuits has resulted in also increased cost of tests. In order to optimize test cost without compromising IC quality, it is imperative to reduce test time. Addressing the issue of prolonged test duration, associated with conventional full-pattern tests approaches, this paper introduces a novel and efficient pattern selection method known as stepwise k-nearest neighbors (KNN). By harnessing the power of a classification model, this method selectively identifies and applies only the most potent patterns, that are likely to lead to test failures. This approach dramatically expedites the testing process, resulting in substantial time savings. Experimental results showcase an impressive 27% enhancement in the accuracy of the proposed approach compared to Support Vector Machine (SVM) techniques. This signifies the effectiveness and superiority of the presented method in enhancing both the efficiency and reliability of IC testing.
Pythagorean fuzzy set, characterized by membership function and non-membership function, has received increasing attention in recent years. In this paper, a new approach to decision making is proposed based on Pythagorean fuzzy preference relation and its additive consistency. Firstly, the concepts of Pythagorean fuzzy preference relations and its additive consistency are introduced, and followed by a discussion of their desirable properties. Then, a linear goal programming model is proposed to determine the consistency of PFPRs. For the PFPRs that does not satisfy the consistency, the consistency index is defined to measure the degree of consistency, and a consistency adjustment algorithm is proposed. Finally, based on the additive consistency, a new algorithm for decision making is presented. An example of CIM(Computing In Memory) is provided, and in comparison with other methods, the validity and rationality of the proposed method are verified.
This paper proposes a novel approach for cascading Hopfield and Hindmarsh-Rose neural networks using memristive synapses. The model integrates two distinct neural networks, each consisting of two identical neurons. First, we design a memristor and analyze its memristive characteristics. We then develop the Hopfield and Hindmarsh-Rose networks with dual neurons, utilizing memristors to couple these sub-networks, forming a new four-neuron neural network. The equilibrium points and system stability are thoroughly analyzed. The dynamic behavior of the network is examined through bifurcation diagrams, two-parameter bifurcation plots, Lyapunov exponents, and phase diagrams, revealing phenomena such as chaos, periodicity, period-doubling bifurcations, and alternating-period chaos. To validate the numerical simulations, we implement the phase diagram on an STM32 hardware platform. Finally, we propose a DNA-based image encryption algorithm leveraging chaotic sequences generated by the network. The performance analysis of encrypted images demonstrates that the algorithm offers strong robustness during transmission and effective resistance against attacks, ensuring privacy and security.
Pedestrian trajectory prediction is a key technical prerequisite for autonomous vehicle trajectory planning. However, a pedestrian is a changeable individual, and their intentions exhibit certain degrees of randomness and uncertainty, which leads to the issue that modeling only past trajectories does not enable the effective description of the random intentions and future trajectory directions of the pedestrian. Therefore, this paper proposes a flexible and embeddable stochastic intention vector construction strategy for modeling sudden pedestrian intention changes in real scenes and for better fitting the stochastic properties of pedestrian behaviors. First, we dynamically fuse historical trajectory information with random factors and construct an intention change probability based on the historical trajectory fitting errors of pedestrians, aiming to explicitly model the associated direction and velocity changes caused by random pedestrian intentions. Second, a new intention loss function is designed to guide the model to adaptively learn the probability of intention changes, which is used to dynamically describe pedestrian intention changes. Our proposed method is generalizable and can be applied as an embeddable module to any baseline pedestrian trajectory prediction method. The experimental results obtained on multiple large-scale public pedestrian trajectory prediction datasets demonstrate that our strategy achieves consistent performance improvements over different baselines.
Locally active memristors with an Edge-of-Chaos kernel (EOCK) represent a significant advancement in the simulation of neuromorphic dynamics. However, current research on memristors with an EOCK remains at the circuit level, without further analysis of their feasibility. In this context, we designed a memristor and installed it in a third-order circuit, where it showed local activity and stability under defined voltage and inductance parameters. This behavior ensured that by varying the input voltage and inductance, the memristor could effectively simulate various neural activities, including inhibitory postsynaptic potential and chaotic waveforms. By subsequently integrating the memristor with an EOCK into a Hopfield neural network (HNN) framework and substituting the self-coupling weight, we observed a rich spectrum of dynamic behaviors, including the rare phenomenon of antimonotonicity bubble bifurcation. Finally, we used hardware circuits to realize these generated dynamic phenomena, confirming the feasibility of the memristor. By introducing the HNN and studying its dynamic behavior and hardware circuit implementation, this study provides theoretical insights into and an empirical basis for developing circuits and systems that replicate the complexity of human brain functions. This study provides a reference for the development and application of EOCK in the future.
Pedestrian trajectory prediction is extremely challenging due to the complex social attributes of pedestrians. Introducing latent vectors to model trajectory multimodality has become the latest mainstream solution idea. However, previous approaches have overlooked the effects of redundancy that arise from the introduction of latent vectors. Additionally, they often fail to consider the inherent interference of pedestrians with no trajectory history during model training. This results in the model’s inability to fully utilize the training data. Therefore, we propose a two-stage motion pattern de-perturbation strategy, which is a plug-and-play approach that introduces optimization features to model the redundancy effect caused by latent vectors, which helps to eliminate the redundancy effects in the trajectory prediction phase. We also propose loss masks to reduce the interference of invalid data during training to accurately model pedestrian motion patterns with strong physical interpretability. Our comparative experiments on the publicly available ETH and UCY pedestrian trajectory datasets, as well as the Stanford UAV dataset, show that our optimization strategy achieves better pedestrian trajectory prediction accuracies than a range of state-of-the-art baseline models; in particular, our optimization strategy effectively absorbs the training data to assist the baseline models in achieving optimal modeling accuracy.